Presentation 2004/12/15
Verifying LVCSR Output at Different Levels with Generalized Posterior Probability
Wai-Kit LO, Frank K. SOONG, Satoshi NAKAMURA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Generalized posterior probability (GPP), a statistical confidence measure, is used for verification of large vocabulary continuous speech recognition (LVCSR) output at subword, word and utterance levels. GPP is obtained by combining exponentially and optimally weighted products of acoustic and language model scores for reappeared units in the reduced search space (e.g., word graph). Experimental results have demonstrated the effectiveness of GPP for verifying LVCSR output at all three levels.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) confidence measure / posterior probability / large vocabulary continuous speech recognition
Paper # NLC2004-84,SP2004-124
Date of Issue

Conference Information
Committee NLC
Conference Date 2004/12/15(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Natural Language Understanding and Models of Communication (NLC)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Verifying LVCSR Output at Different Levels with Generalized Posterior Probability
Sub Title (in English)
Keyword(1) confidence measure
Keyword(2) posterior probability
Keyword(3) large vocabulary continuous speech recognition
1st Author's Name Wai-Kit LO
1st Author's Affiliation Spoken Language Translation Research Laboratories, ATR()
2nd Author's Name Frank K. SOONG
2nd Author's Affiliation Spoken Language Translation Research Laboratories, ATR
3rd Author's Name Satoshi NAKAMURA
3rd Author's Affiliation Spoken Language Translation Research Laboratories, ATR
Date 2004/12/15
Paper # NLC2004-84,SP2004-124
Volume (vol) vol.104
Number (no) 540
Page pp.pp.-
#Pages 6
Date of Issue